BUG#21542698 - LOAD DATA INFILE: ASSERTION FAILED: ML == 2 || ML == 4

This is due to a too strict assertion, which could not handle the invlaid
characters. We just have to accept the invalid characters in the assertion.

Reviewed-by: Roy Lyseng <Roy.lyseng@oracle.com>

Author: Bin Su

include/m_ctype.h

@@ -764,7 +764,8 @@ uint my_mbcharlen_ptr(const CHARSET_INFO *cs, const char *s, const char *e);
   @param[in] a first byte of gb18030 code
   @param[in] b second byte of gb18030 code
   @return    the length of gb18030 code starting with given two bytes,
-             the length would be 2 or 4
+             the length would be 2 or 4 for valid gb18030 code,
+             or 0 for invalid gb18030 code
 */
 #define my_mbcharlen_2(s, a, b)       ((s)->cset->mbcharlen((s),((((a) & 0xFF) << 8) + ((b) & 0xFF))))
 /**

mysql-test/r/ctype_gb18030.result

@@ -1795,6 +1795,10 @@ CCE7	1
 99CC	1
 8BF5819AEDC3	1
 DROP TABLE t1;
+CREATE TABLE t1 (a int) ENGINE=InnoDB;
+LOAD DATA INFILE '../../std_data/bug21542698.dat' INTO TABLE t1 CHARACTER SET gb18030;
+ERROR HY000: Invalid gb18030 character string: 'MZ'
+DROP TABLE t1;
 DROP DATABASE gb18030;
 SET sql_mode = default;
 #

mysql-test/std_data/bug21542698.dat

@@ -258,6 +258,12 @@ INSERT INTO t1 VALUES (_gb18030 0x8BF5819AEDC3), (_gb18030 0x99CC), (_gb18030 0x
 SELECT ANY_VALUE(HEX(c)), COUNT(c) FROM t1 GROUP BY c COLLATE gb18030_chinese_ci;
 DROP TABLE t1;
 
+# Test if LOAD DATA with gb18030 chars can handle invalid chars
+CREATE TABLE t1 (a int) ENGINE=InnoDB;
+--error ER_INVALID_CHARACTER_STRING
+LOAD DATA INFILE '../../std_data/bug21542698.dat' INTO TABLE t1 CHARACTER SET gb18030; 
+DROP TABLE t1;
+
 DROP DATABASE gb18030;
 SET sql_mode = default;
 --echo #

mysql-test/t/ctype_gb18030.test

@@ -1568,10 +1568,11 @@ READ_INFO::~READ_INFO()
       if (chr1 != my_b_EOF)                                                   \
       {                                                                       \
         len= my_mbcharlen_2((cs), (chr), chr1);                               \
-        /* Must be gb18030 charset */                                         \
-        DBUG_ASSERT(len == 2 || len == 4);                                    \
+        /* Character is gb18030 or invalid (len = 0) */                       \
+        DBUG_ASSERT(len == 0 || len == 2 || len == 4);                        \
       }                                                                       \
-      PUSH(chr1);                                                             \
+      if (len != 0)                                                           \
+        PUSH(chr1);                                                           \
     }                                                                         \
   } while (0)